Identification and nanomechanical characterization ofthe fundamental single-strand protofilaments ofamyloid α-synuclein fibrilsFrancesco Simone Ruggeria,b,1, Fabrizio Benedettic, Tuomas P. J. Knowlesb,d, Hilal A. Lashuele, Sergey Sekatskiia,and Giovanni Dietlera,1

aLaboratory of Physics of Living Matter, École Polytechnique Fédérale de Lausanne (EPFL), 1015 Lausanne, Switzerland; bDepartment of Chemistry,University of Cambridge, CB2 1EW Cambridge, United Kingdom; cCentre for Integrative Genomics, Faculty of Biology and Medicine, University of Lausanne,1015 Lausanne, Switzerland; dCavendish Laboratory, Department of Physics, University of Cambridge, CB3 0HE Cambridge, United Kingdom;and eLaboratory of Molecular and Chemical Biology of Neurodegeneration, Brain Mind Institute, Faculty of Life Sciences, École Polytechnique Fédérale deLausanne (EPFL), 1015 Lausanne, Switzerland

Edited by Ehud Gazit, Tel Aviv University, Tel Aviv, Israel, and accepted by Editorial Board Member F. Ulrich Hartl May 23, 2018 (received for review December6, 2017)

The formation and spreading of amyloid aggregates from thepresynaptic protein α-synuclein in the brain play central roles inthe pathogenesis of Parkinson’s disease. Here, we use high-resolution atomic force microscopy to investigate the early oligo-merization events of α-synuclein with single monomer angstromresolution. We identify, visualize, and characterize directly the small-est elementary unit in the hierarchical assembly of amyloid fibrils,termed here single-strand protofilaments. We show that protofila-ments form from the direct molecular assembly of unfolded mono-meric α-synuclein polypeptide chains. To unravel protofilaments’internal structure and elastic properties, we manipulated nano-mechanically these species by atomic force spectroscopy. The single-molecule scale identification and characterization of the fundamentalunit of amyloid assemblies provide insights into early eventsunderlying their formation and shed light on opportunities fortherapeutic intervention at the early stages of aberrant proteinself-assembly.

amyloid | protein aggregation | atomic force microscopy | early molecularassembly | force spectroscopy

The aggregation of α-synuclein into cross–β-sheet fibrillarstructures plays a central role in the neuropathology of

Parkinson’s disease (PD) and other neurodegenerative diseases(Synucleinopathies) (1–6). Despite the widespread association ofα-synuclein aggregation with the formation of Lewy bodies, thebrain pathology that characterizes PD, the mechanistic roles ofα-synuclein oligomerization and fibril formation in cell death anddisease onset have remained challenging to identify (7). In par-ticular, the nature of the specific state(s) of proteins that causecytotoxicity is not well established. It is clear, however, that ag-gregated rather than monomeric forms of α-synuclein areprimarily implicated in α-synuclein–mediated toxicity andpathology spreading in PD and related Synucleinopathies (3).The kinetics of α-synuclein aggregation follows a nucleation-

dependent process, where monomers initially aggregate intooligomeric species leading to the final formation of the amyloidfibrillar aggregates (8–10). The kinetics and rates of α-synucleinaggregation, as well as the morphologies of the different aggre-gates it populates on the pathway to amyloid formation, arestrongly influenced by the experimental conditions used. Inparticular, incubation under shaking condition or in the presenceof lipids is known to accelerate the fibril formation process(4, 11). Increasing evidence suggests that the prefibrillar aggre-gates, such as oligomers and protofibrils, or the process offibrillization, rather than the final mature fibrillar products, areresponsible for α-synuclein toxicity in vivo (12–14).Although the ability to reproduce α-synuclein fibril formation

in vitro has advanced our understanding of the molecular and

structural determinants of α-synuclein aggregation and pathologyonset, our understanding of early oligomerization events and themode and strength of monomer interactions at different stages ofthe amyloid pathway remains very limited. For these reasons, theidentification and characterization of aggregation intermediateson the pathway to α-synuclein amyloid formation is fundamentalto understanding the mechanisms of amyloid formation, identi-fication of the toxic species, and the development of novel di-agnostic and therapeutic strategies for PD and related disorders.Atomic force microscopy (AFM) and single-molecule force

spectroscopy represent important research tools for achievingquantitative understanding of the mechanism of amyloid for-mation and characterization of different species on the amyloidpathway at the single-molecule scale (15–19). Several AFMstudies suggested that many amyloidogenic proteins follow ahierarchical assembling model in which a certain number ofelongated prefibrillar species with nanometer-size diameters andmicrometer-scale lengths, termed protofilaments, intertwine or

Significance

Today, more than 40 million people worldwide are affected byneurodegenerative disorders. These diseases are associated atthe molecular level with the aggregation of specific proteinsinto insoluble fibrils, termed amyloids. Increasing evidencesuggests that the intermediate aggregates, rather than the fi-nal fibrillar products, are implicated in toxicity in vivo. How-ever, the investigation of the conversion of proteins intoamyloids represents a formidable experimental challenge be-cause of their nanoscale and heterogeneous nature. Here, wereport the identification of a mechanism of early assembly ofmonomeric proteins into elongated intermediates during am-yloid formation. The biophysical characterization of novel in-termediate molecular species is fundamental to unravel theirmechanism of formation and gain insights into their potentialtoxicity and role in pathology.

Author contributions: F.S.R., T.P.J.K., H.A.L., S.S., and G.D. designed research; F.S.R. andF.B. performed research; F.S.R., F.B., and T.P.J.K. analyzed data; and F.S.R., F.B., T.P.J.K.,H.A.L., S.S., and G.D. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission. E.G. is a guest editor invited by theEditorial Board.

This open access article is distributed under Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND).1To whom correspondence may be addressed. Email: fsr26@cam.ac.uk or giovanni.dietler@epfl.ch.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1721220115/-/DCSupplemental.

Published online June 25, 2018.

7230–7235 | PNAS | July 10, 2018 | vol. 115 | no. 28 www.pnas.org/cgi/doi/10.1073/pnas.1721220115

Dow

nloa

ded

by g

uest

on

Janu

ary

24, 2

021

laterally associate to form the final mature amyloid fibrillarstructures with a typical diameter in the order of 6–13 nm; thiswas also the case with α-synuclein (5, 20–26). Previous studiessought to quantify the structural, nanomechanical properties andthe interaction forces of the protofilaments within matureamyloid fibrils (15–17). These studies demonstrated how singleaggregate imaging and manipulation methods enable studyingthe mechanics and structural dynamics of amyloid inner prop-erties and formation. However, to date, it has not been possibleto isolate and characterize the earliest elementary units in theassembly process formed from a single beta-sheet. This difficultyhas originated in large part from their transient nature and theirexpected very small dimensions with a thickness correspondingto a single strand. Thus, our current knowledge of their structuralproperties is based primarily on investigating their structure andmorphology when already present within mature fibrils.Here, we use high-resolution AFM and force spectroscopy to

revisit the pathway of α-synuclein fibril formation, with specialemphasis on early molecular assembly events, to identify andcharacterize the elemental protofilament unit leading to the for-mation of mature fibrillar amyloids. We report the detection andcharacterization of the smallest elementary unit in the hierarchicalassembly of amyloid fibrils of α-synuclein, which assemblesdirectly from unfolded monomers termed here single-strandprotofilaments. These protofilaments possess a subnanometer-scale cross-sectional diameter which is smaller than previouslyobserved aggregated species of α-synuclein. Indeed, their sub-nanometer diameter is smaller than the typical dimensions ofpreviously observed elongated nonmature aggregated species of

α-synuclein [diameter of about 3–6 nm (5)], much smaller thanthe typical dimensions of mature amyloid fibrils (diameter of about6–13 nm), and have a radius of gyration smaller than an individualmonomer itself (∼3 nm) (5, 24). The application of a scale rea-soning demonstrates that the single-strand protofilament speciesare directly assembled from monomers of α-synuclein, and with aheight corresponding to a single strand. Finally, we performedatomic force spectroscopy studies to characterize directly the innermechanical, structural, and thermodynamical properties of single-strand protofilaments by means of statistical theory of biopolymers.We show that when subjected to mechanical dissection, the single-strand protofilaments unravel into components showing an averagepersistence length consistent with the elastic properties of a poly-mer chain composed of unstructured monomers of α-synuclein, aspredicted by their cross-sectional dimensions.

Results and DiscussionAggregation of α-Synuclein Proceeds Through Formation ofProtofilament Aggregates. To unravel the process of α-synucleinfibrillization at the single-molecule scale, we exploited AFMhigh-resolution 3D morphology imaging with the deposition of thenegatively charged α-synuclein protein on both an atomically flatnegative mica and on a positively functionalized mica substrate(mica-APTES) (Fig. 1 and SI Appendix, Fig. S1) (27). We in-cubated freshly filtered (100 kDa) α-synuclein monomeric solutionin vitro under shaking conditions at 37 °C for 10 d. We collectedan aliquot of the solution at 0, 1, and 10 d, and then we depositedan aliquot of each sample on the substrate to acquire morphologymaps of the sample by AFM. The sample was monitored in parallel

0

10

20

30

40

Cou

nt

monomers

dimers

Protofilaments

Singlestrand

Protofibrils

Fibrils

0 2 4 6 8 100

10

20

30

Cou

nt

Height (nm)

0.0 0.4 0.8 1.20

10

20

Cou

nt

Singlestrand

Doublestrand

Dimers

Monomers

ProtofibrilsFibrils

protofilaments higher order aggregates fibrilsmonomers0.4 nm 0.8 –6 nm 6 –10 nm

Single Strandcross-section

0.4 nm

Molecular assembly Hierarchical assembly

Double Strandcross-section

nm

nm

0.0 0.4 0.80.1 μm

Height (nm)

Height (nm)

nm

nm

A

B

C

D

Fig. 1. α-Synuclein fibrillization under shaking con-ditions. (A–C) Aggregation time course monitored byhigh-resolution AFM imaging, detail of the cross-sectional height of a species and single-molecule statis-tical analysis, with histogram distributions of the cross-sectional height of the species observed at (A) 0, (B) 1,and (C) 10 d (Z scales are in nanometers). (D) Model ofα-synuclein fibril formation showing the (i) initial mo-lecular assembly leading to the formation of the newlyidentified single-strand protofilaments (in light red) and(ii) subsequent hierarchical assembly leading to theformation of mature amyloid fibrils. At the bottom, thetypical AFM height of each species is indicated.

Ruggeri et al. PNAS | July 10, 2018 | vol. 115 | no. 28 | 7231

BIOPH

YSICSAND

COMPU

TATIONALBIOLO

GY

BIOCH

EMISTR

Y

Dow

nloa

ded

by g

uest

on

Janu

ary

24, 2

021

by circular dichroism and thioflavin T to independently confirm thestate of aggregation in bulk, as we previously reported (27, 28).Before incubation and immediately after the dissolution of

α-synuclein in buffer, only monomers and early oligomers werepresent on the surface and no fibrillar species were visible (Fig.1A and SI Appendix, Fig. S2) (27, 28). A single-molecule statis-tical analysis revealed that the smallest oligomeric populationhad a cross-sectional height of 0.3–0.4 nm (Fig. 1A). The heightof this first population corresponded to the height on a surface ofa monomeric polypeptide chain, as previously measured for anunfolded or partially folded protein by scanning probe microscopy(27, 29, 30). Thus, the first oligomeric population corresponded tomonomeric α-synuclein (SI Appendix). In the histogram distributionof height of the early oligomeric populations, we also observedanother population with double the cross-sectional height of themonomeric species, which we associated to dimeric species. After24 h of incubation, we could observe the formation of elongatedprotofilaments species having subnanometer-scale diameter andmicrometer-scale length (Fig. 1B). We observed two polymorphs ofthe protofilament species. The first species had a cross-sectionalheight of 0.3–0.4 nm, corresponding to the height of monomericα-synuclein. The second species had a cross-sectional height of 0.7–0.9 nm, which corresponded to the cross-sectional height of dimers(Fig. 1 A and B). Remarkably, we reported the observation ofsubnanometer fibril-like aggregates of α-synuclein. We termed theirelemental unit, possessing the smallest cross-sectional diameter of asingle monomer (0.3–0.4 nm), as single-strand cross-section proto-filaments or more easily single-strand protofilaments (Fig. 1B).Considering that the second polymorphic population had doublethe height of single-strand protofilaments, we termed them double-strand cross-section protofilaments. Finally, after 10 d of incubation,we observed the formation of higher order aggregates and matureamyloid fibrils of α-synuclein with height ranging between 4–6 and6–9 nm, respectively (Fig. 1C). This observation suggested that theprotofilaments species were on-pathway of the formation of maturefibrils [i.e., height by AFM of 6–9 nm (24)]. As depicted in Fig. 1D,our results suggested that the early aggregation of α-synucleinproceeded through the molecular assembly of monomers into thesingle-strand protofilaments, which then interact and assemblehierarchically to form higher order aggregates leading to the forma-tion of mature amyloid fibrils.Under shaking conditions, α-synuclein fibrillization occurred

within 10 d, whereas several months of incubation are required toachieve complete fibrillation of α-synuclein under quiescent con-ditions (4, 27, 28). To investigate the process of protofilamentformation with higher time resolution, we studied the aggregationof α-synuclein under quiescent conditions at 37 °C, over a longperiod of 181 d (Fig. 2). To avoid any bias for aggregation in ourexperimental setup, at each time point, we deposited an aliquot, ofα-synuclein solution aggregated in vitro, on two substrates with theopposite state of charge: negative bare mica and positive mica-APTES (SI Appendix, Fig. S3). Analysis of the freshly preparedα-synuclein samples showed that they contain predominantly mo-nomeric protein, and some oligomeric species were also observed(Fig. 2A). In static conditions, protofilament species were firstobserved after 2 months (Fig. 2B, 60 d) of incubation in vitro.Then, they elongated and increased in number and length after6 months of incubation (Fig. 2C, 181 d). The kinetics of aggregationwas confirmed independently on bare mica and functionalizedmica-APTES substrates (Fig. 2C and SI Appendix, Fig. S3). Fur-thermore, we independently detected the protofilaments, after180 d of incubation, by transmission electron microscopy (TEM)(Fig. 2D). AFM high-resolution imaging enabled observing that thedouble-strand protofilaments formed from the association of in-dividual single-strand protofilaments (Fig. 2E). Similarly to shakingconditions, a statistical characterization of the cross-sectional di-mensions of individual protofilament species showed that theycould be separated in two main families according to their AFM

height. The first family was corresponding with the single-strandprotofilaments, with an average height of 0.3–0.4 nm, and thesecond family was corresponding with the double-strand proto-filaments, which had an average height of 0.7–0.8 nm (Fig. 2F). Theslow speed of aggregation in static conditions enabled monitoring,over a long period of time, the relationship between the height ofmonomers and dimers with the height of single- and double-strandprotofilaments (Fig. 2 G–I). Independently from the incubationtime, similarly as observed at time 0 d, the smallest oligomericpopulation possessed a cross-sectional height of ∼0.3–0.4 nm, whilethe second smallest dimeric population had a cross-sectional heightof ∼0.6–0.9 nm (Fig. 2G). Correspondingly, within the time periodinvestigated here, the two protofilament families showed identicalcross-sectional height, of ∼0.4 nm and 0.8 nm, respectively. To ruleout any possible effect of the surface of deposition on the mor-phological properties of the protofilaments, we deposited thesample solution on the two substrates with the opposite state ofcharge, negative mica (Fig. 2H) and positive mica-APTES (Fig. 2I).The protofilaments showed identical properties on mica and mica-APTES (SI Appendix, Fig. S3). Thus, the surface of deposition didnot affect the structural properties of the protofilaments.The AFM imaging combined with a single-molecule statistical

analysis of the average heights of oligomeric and protofilamentsspecies at different time points (Fig. 2 G–I) demonstrate that asingle-strand protofilament possesses the same cross-sectionalheight of a monomer (orange dashed line), whereas a double-strand protofilament, formed from two single strands interacting,has the same cross-sectional height of the dimeric population(red dashed line).

Single-Strand Protofilaments Are Assembled Monomeric Chains. Inboth quiescent and shaking incubation conditions, the height ofsingle-strand protofilaments corresponded to the height of a

0 0.3 0.7 1.00

100

Double Strand

Single Strand

Singlestrand

Double strand

0 50 100 150 2000.00.20.40.60.81.01.2 Monomers

Dimers

Heig

ht(n

m)

Days (d)0 50 100 150 2000.0

0.20.40.60.81.01.2 Single strand

Double strand

Days (d)0 50 100 150 2000.0

0.20.40.60.81.01.2

Days (d)

mica

50 nm

0

1

0

1

0

1

2

3

4

50 nm 50 nm

20 nm

mica

Single strandDouble strand

mica-APTES

Heig

ht(n

m)

Heig

ht(n

m)

50 nmHeight (nm)

A B C

D E F

G H I

Fig. 2. Slow aggregation in quiescent conditions of monomeric α-synucleininto protofilament species. AFM image of (A) monomers and oligomers cov-ering the surface at 0 d, (B) protofilaments after 60 d and (C–E) after 181 d asobserved by (C) AFM and (D) TEM. (E) Detail of intertwining protofilaments.(F) Single-molecule statistical analysis and histograms of the cross-sectionalheight of protofilaments after 181 d. Comparison of the cross-sectionalheight of (G) the smallest oligomeric populations, corresponding to mono-mers and dimers, with the cross-sectional height of protofilaments on (H) baremica and (I) mica-APTES as a function of the incubation time in vitro. (Z scalesare in nanometers; error bar is 1 SD.)

7232 | www.pnas.org/cgi/doi/10.1073/pnas.1721220115 Ruggeri et al.

Dow

nloa

ded

by g

uest

on

Janu

ary

24, 2

021

single polypeptide chain. We measured an average height of 0.3–0.4 nm for both monomers and single-strand protofilaments(Figs. 1 and 2 and SI Appendix, Fig. S2). The measurement ofheight by AFM, in air for a relaxed object on a surface smallerthan a couple of nanometers in height, could be affected byunderestimation (SI Appendix) (31, 32). Thus, we deposited ourprotofilament aggregates together with circular DNA, which wasused as an internal ruler for the measurement of the protofila-ment’s real height (Fig. 3A). Indeed, DNA structure has beenextensively studied and it is universally accepted as a flexibledouble helix with a ∼2-nm diameter. AFM measurements ofDNA, in the same scanning conditions as for the protofilamentimaging, showed a measured height of ∼0.8–1 nm (Fig. 3B),consistent with previous reports (32, 33). Single-strand structurespossessed almost one half of the measured height of DNAmolecules, whereas double-strand protofilaments had nearly thesame height as DNA (Fig. 3B). Supposing that DNA and theaggregates have a similar geometrical cross-section shape, we couldconclude that the real diameter of single-strand protofilaments wasof ∼0.6–1 nm. The height of single-strand protofilaments was muchsmaller than the radius of gyration of a single monomer in globularconformation (2–3 nm) and of the order of the experimentallymeasured height of a single polypeptide chain (30). In addition, theprotofilament species had a much smaller dimension of anypreviously observed aggregated species of α-synuclein. In fact,their subnanometer diameter is smaller than the typical dimensionsof previously observed elongated nonmature (4–6 nm height) andmature (6–10 nm) amyloid aggregated species of α-synuclein (Fig.1D) (5, 24). Hence, the subnanometer cross-sectional height ofsingle-strand protofilaments excluded the possibility that more thanone monomer was present within their cross-section. Besides, theprotofilament’s length, in the order of hundreds of nanometers,well exceeds the one of a fully extended monomeric chain ofα-synuclein (∼50 nm, 0.35 nm for amino acid). Thus, we concludedthat the protofilaments consisted of assembled polypeptide chainsof monomers of α-synuclein in series.

Nanomechanical Manipulation of Protofilaments by Atomic ForceSpectroscopy. The small cross-sectional diameter of protofilamentspecies (<1 nm) did not permit the use of the well-establishedsingle-molecule methods, such as AFM-based nanoindentationand infrared nanospectroscopy, to probe their biophysicalproperties at the single-molecule scale (34, 35). For this reason, wepursued a nanomechanical manipulation of the protofilaments bymeans of atomic force spectroscopy (Fig. 4). The tip of our AFMcantilever was approached against the substrate, where proteinaggregates were deposited. Then, it was retracted and force ex-tension curves were recorded. The investigation of the spectro-scopic force response for oligomeric and protofilament speciesshowed a remarkable difference (SI Appendix, Fig. S4). At time 0 d,in the force spectra, there was no presence of any complex or

repeatable force response. Conversely, after 181 d of incubation,highly reproducible multiple force plateaus (Fig. 4 and SI Appen-dix) forming a staircase-like picture were observed in the force–distance curves. These constant force-pulling events had a muchlonger length (up to 400 nm) than a fully extended α-synucleinmonomeric polypeptide chain (∼50 nm). Therefore, the forceplateaus were due to the presence of the two types of protofila-ments on our surface, and as in previous studies, we associated theconstant force-pulling events to the detaching of the protofilamentsfrom the surface (15, 16). We present in Fig. 4A a model of thesimplest possible detachment event from the surface (15–17). Ini-tially, the AFM tip was approached on the substrate and it wascapable of breaking a protofilament (Fig. 4 A, 1). Indeed, as al-ready measured and clearly demonstrated in the case of matureamyloid structures, which are notably stiffer than monomericstructures, ∼500 pN were already sufficient to break by indentationthe aggregate (36). Then, the tip was moved far from the surface,and we started pulling the protofilament and detaching it from thesubstrate (Fig. 4 A, 2) until the nanomechanical contact is lost (Fig.4 A, 3). In addition to the detachment of a single protofilament,several mechanical responses were measured during the process ofprotofilaments pulling from the surface (SI Appendix, Figs. S4 andS5). We focused our attention on the force plateaus (Fig. 4 B–D),which constituted the majority of the events (89%) compared withthe nonspecific effects (11%). More specifically, we characterizedthe force plateau’s length and height distribution. We performed astatistical analysis of the plateaus’ lengths to prove that force pla-teaus were shorter than the actual protofilament’s length (SI Ap-pendix, Fig. S6). The force plateaus possessed an average length of∼70 nm (Fig. 4E), whereas the protofilaments had a length with anaverage value of ∼170 nm (Fig. 4F). Then, to quantify the processof protofilaments’ detachment from the surface and to obtain anexperimental estimation of the forces involved, we measured theheight of the observed plateau force steps. We measured the force

DNA SingleStrand

Doublestrand

0.0

0.2

0.4

0.6

0.81.0

Hei

ght [

nm]

(25%~75%) (25%~75%) (25%~75%) Mean ± 1 SD Median Line

Circular DNA Protofilaments

A B

Fig. 3. Calibration of single-strand protofilament height relative to circulardouble-strand DNA. (A) AFM image of protofilaments and circular DNA (Zscale is in nanometers). (B) Scatter plot and comparison of the height ofcircular DNA and protofilaments.

ΔF2 3

0 100 200 300 4000

140

Coun

t

Plateau Length (nm)0 100 200 300 4000

18

Coun

t

Protofilament Length (nm)

Single strand Double strand

1

2 3 41

23 4

A

B C D

E F

Fig. 4. Force responses of manipulated α-synuclein protofilaments. (A) Modelof single protofilament’s pulling and relative force–distance curve. (1) AFM tipis strongly pressed against a protofilament. (2) The strong pressure causespolymer breaking and a constant force plateau due to its unzipping from thesurface. (3) Tip–filament contact is lost during the pulling. (4) Tip return to theinitial position. (B) Single force plateau due to the pulling of one protofila-ment. (C) Double plateau due to the pulling of two protofilaments. (D) Pullingof four protofilaments. (E) Distribution of the lengths of constant force pullingevents (plateaus). (F) Distribution of the lengths of protofilaments.

Ruggeri et al. PNAS | July 10, 2018 | vol. 115 | no. 28 | 7233

BIOPH

YSICSAND

COMPU

TATIONALBIOLO

GY

BIOCH

EMISTR

Y

Dow

nloa

ded

by g

uest

on

Janu

ary

24, 2

021

difference ΔF between the average height of a plateau and thebaseline in the force curve (Fig. 4B). A distribution plot of thisquantity showed a multimodal distribution with four peaks andsome more rare higher force events, which we considered gener-ically as the fifth peak (SI Appendix, Fig. S7A). The force peaksappeared at integer multiples of a single detachment event valueof ∼50 pN (SI Appendix, Fig. S7B). We could extract an averagequantized value of detachment force from the substrate of 57 ±7.5 pN (SI Appendix, Fig. S7 C and D). Since single-strand pro-tofilaments were longer and more abundant than double-strandones (Fig. 4F), the pulling or unzipping of double-strand proto-filaments could bring only a negligible broadening of the forcedistribution peaks and could be neglected (SI Appendix).The quantized values of force plateaus obeyed a linear law,

demonstrating that multiple plateaus were linked to the multipledetachments of several protofilaments from the surface. AMonte Carlo simulation of the process, modeling the breaking ofn parallel bonds from the surface, explained to good approxi-mation the essential features of this quantization and reproducedthe measured experimental values, thus confirming that thepresence of n plateaus was due to the nanomechanical de-tachment from the surface of n protofilaments in parallel (SIAppendix, Fig. S7 C and D).

Single-Strand Protofilaments Have Disordered Structural Organization.Remarkably, we observed two classes of force plateaus. The firstclass were flat constant force curves, while the second class wereconstant force plateaus with stretching events, which were ob-served inside or at the border of a plateau (SI Appendix, Fig. 6 A–C). Thus, after the characterization of the detachment process ofthe plateaus, we focused our attention on the plateaus presentingstretching events.The quantization of the force plateaus (SI Appendix, Fig. S7)

enabled us to exclude any internal unzipping of the protofilaments;thus, we concluded that during the pulling the protofilamentspecies were simply manipulated and detached from the surface.This observation was in agreement with the above demonstratedconclusion (Figs. 1–3) that single-strand protofilaments have cross-sectional dimensions just of the order of the carbon–nitrogenbackbone of a single α-synuclein monomer. The typical value ofthe length of the stretching events was of 15 ± 10 nm; thestretching events were in all cases <40 nm (n = 35) and smallerthan the maximal extension of a single polypeptide chain ofα-synuclein (∼50 nm; 0.35 nm for the amino acid). Thus, we couldassociate the stretching events to the elastic deformation of thecarbon–nitrogen backbone of a single monomeric-like chain insidea protofilament (Fig. 5D). We investigated the nature of thestretching events by means of the statistical theory of biopolymersto retrieve the internal mechanical properties of the monomerswithin the aggregates (37). By fitting the stretching events with aworm-like chain model (WLC; see SI Appendix, Eq. S2), we coulddetermine an internal average persistence length A of the single-strand protofilaments of A = 0.29 ± 0.16 nm (n = 35). This valuewas fully consistent with the persistence length of an unfoldedprotein (38). It was also already measured by AFM force spec-troscopy that, upon similar stretching by AFM, mature amyloidfibrils formed by the glucagon peptide have an average internalpersistence length of 0.70 ± 0.15 nm (17). Remarkably, the internalpersistence length measured for mature fibrils was more than twiceas big as in the case of the single-strand protofilaments.Furthermore, we exploited the force curves in Fig. 5 A–C to

measure the thermodynamic free energy change upon stretchingof an α-synuclein monomer within the single-strand protofila-ments. First, we measured the irreversible work (W) needed tostretch and deform the polypeptide chain. The integration of thearea below the stretching events enabled us to measure an av-erage value of W = 350 ± 100 kJ/mol. Then, we calculated theentropic term of the free energy necessary to separate the ends

of the flexible chain by a distance z, as previously reported in theliterature (37, 39). We obtained an average value of FE = 40 ±20 kJ/mol. We considered the irreversible work W done on thesystem the upper limit of the free energy change of protofila-ments upon stretching, while we considered the entropic part ofthe free energy FE the minimal free energy change upon theextension of the monomeric chain. Thus, when submitted to ir-reversible stretching, we could estimate that the free-energychange of a single-strand protofilament ranges between approxi-mately F ∼ 40 and 350 kJ/mol. As expected, the irreversible workdone to stretch a monomer is higher than its change of entropicfree-energy FE upon extension. Since the extended conformationof the polypeptide chain of a monomer does not allow stable hy-drogen bonding, the irreversible work done to extend the chain canbe likely related to the breaking of intramolecular hydrogen bonds(30, 40–42). We estimate that the energy to extend the monomericchain to break a single hydrogen bond is given by EH =F=n, whereF is the free energy and n the number of intramolecular hydrogenbonds. For simplicity, we consider the presence of 1 hydrogen bondfor each amino acid (41, 43). For a total free energy of 350 kJ/moland a stretching length of 15 nm, we obtained an average estimatedvalue of the free energy to break a hydrogen bond of EH = 8 ±2 kJ/mol. This value is in excellent agreement with the reportedvalue in the literature of the free energy necessary to breakhydrogen bonding in a monomeric protein (44–47). Furthermore,the measured strength of hydrogen bonding in a monomer withinthe protofilament is remarkably smaller than the energy of an in-termolecular hydrogen bond in amyloid fibrils (∼15–30 kJ/mol)(48–51).In conclusion, the single-strand protofilament aggregates do

not possess the typical mechanical and thermodynamic proper-ties of mature amyloid fibrils and possess a less-rigid structurearising by the early molecular assembly of unfolded monomers.

ConclusionsIn this study, we demonstrate that the early aggregation ofα-synuclein first proceeds through the assembly of monomers intoelongated chain-like aggregates with a cross-section of a singlestrand. These aggregated species are composed of a single chain ofmonomeric α-synuclein in series (single strand) and possess sub-nanometer diameter and up to micrometer scale length. Our resultsdemonstrate that these intermediate species are the smallestobservable elementary unit in the hierarchical assembly of amyloidfibrils. It is widely accepted that the intermediate products of theaggregation rather than the final mature fibrillar products could be

Forc

e

Distance

Stretching

Stretching

A B C

D

Fig. 5. Characterization internal nanomechanical properties of protofila-ments through stretching. (A) Constant force plateau with a stretching eventinside, (B) constant force plateau with two stretching events inside, and (C)force plateau preceded by a stretching event, fitted by the WLC model. (D)Schematic representation of protofilament internal stretching.

7234 | www.pnas.org/cgi/doi/10.1073/pnas.1721220115 Ruggeri et al.

Dow

nloa

ded

by g

uest

on

Janu

ary

24, 2

021

responsible for α-synuclein toxicity in vivo (8). “Oligomers” is a genericterm to indicate a heterogeneous ensemble of amyloid intermediates,which are normally depicted as spheroidal or toroidal objects.Several reports in the literature indicate that size, shape, andstructure of these species are strictly related to their toxicity(6, 52, 53). The understanding of the early molecular assemblyof α-synuclein and the identification and characterization of anew elongated oligomeric intermediate is of fundamental impor-tance, since its different biophysical properties could be related todifferent mechanisms of amyloid toxicity in vivo and could offernew opportunities for developing new therapeutic strategies targetingthe early stages of the assembly process.We describe an experimental approach that would enable

investigating and comparing the interaction forces of amyloidaggregates on different surfaces, such as hydrophobic ones.Unraveling these forces is central to understanding the mechanismof amyloid fibril formation and its role in the pathogenesis of PD

and Synucleinopathies. Finally, the values obtained for the mo-lecular interaction forces within and between the protofilamentsshed light on the thermodynamics of the early hierarchical amyloidassembling process, and could furthermore provide cues to unrav-eling the fibrillation process at surfaces, such as cell membranes.

Materials and MethodsDetailed information on the experimental methods can be found in the SIAppendix. α-Synuclein was produced in Escherichia coli and then purified bypreviously accepted protocols (54). AFM images and force spectroscopy curveswere realized using a Nanoscope IIIa (Bruker Corp.) and NX10 (Park Systems).TEM imaging was performed on a Tecnai Spirit BioTWIN electron microscope.

ACKNOWLEDGMENTS. This work was supported by the Darwin College; theSwiss National Foundation for Science Grants 152958, P300P2_171219, and200021_162767; and the European Research Council under the EuropeanUnion's Seventh Framework Programme (FP7/2007-2013) through the ERCgrant PhysProt (Agreement 337969).

1. Goedert M (1997) Familial Parkinson’s disease. The awakening of alpha-synuclein.Nature 388:232–233.

2. Conway KA, Harper JD, Lansbury PT (1998) Accelerated in vitro fibril formationby a mutant alpha-synuclein linked to early-onset Parkinson disease. Nat Med 4:1318–1320.

3. Uversky V, Eliezer D (2009) Biophysics of Parkinsons disease: Structure and aggrega-tion of-synuclein. Curr Protein Pept Sci 10:483–499.

4. Giehm L, Lorenzen N, Otzen DE (2011) Assays for α-synuclein aggregation. Methods53:295–305.

5. Khurana R, et al. (2003) A general model for amyloid fibril assembly based on mor-phological studies using atomic force microscopy. Biophys J 85:1135–1144.

6. Lashuel HA, Lansbury PT, Jr (2006) Are amyloid diseases caused by protein aggregatesthat mimic bacterial pore-forming toxins? Q Rev Biophys 39:167–201.

7. Lansbury PT, Lashuel HA (2006) A century-old debate on protein aggregation andneurodegeneration enters the clinic. Nature 443:774–779.

8. Knowles TPJ, Vendruscolo M, Dobson CM (2014) The amyloid state and its associationwith protein misfolding diseases. Nat Rev Mol Cell Biol 15:384–396.

9. Tuttle MD, et al. (2016) Solid-state NMR structure of a pathogenic fibril of full-lengthhuman α-synuclein. Nat Struct Mol Biol 23:409–415.

10. Bloch DN, Miller Y (2017) Study of molecular mechanisms of α-synuclein assembly:Insight into a cross-β structure in the N-termini of new α-synuclein fibrils. ACS Omega2:3363–3370.

11. Flagmeier P, et al. (2016) Mutations associated with familial Parkinson’s disease alterthe initiation and amplification steps of α-synuclein aggregation. Proc Natl Acad SciUSA 113:10328–10333.

12. Lashuel HA, Overk CR, Oueslati A, Masliah E (2013) The many faces of α-synuclein:From structure and toxicity to therapeutic target. Nat Rev Neurosci 14:38–48.

13. Hong DP, Han S, Fink AL, Uversky VN (2011) Characterization of the non-fibrillarα-synuclein oligomers. Protein Pept Lett 18:230–240.

14. Mahul-Mellier AL, et al. (2015) Fibril growth and seeding capacity play key roles inα-synuclein-mediated apoptotic cell death. Cell Death Differ 22:2107–2122.

15. Kellermayer MS, et al. (2005) Reversible mechanical unzipping of amyloid beta-fibrils.J Biol Chem 280:8464–8470.

16. Karsai A, et al. (2006) Mechanical manipulation of Alzheimer’s amyloid beta1-42 fi-brils. J Struct Biol 155:316–326.

17. Dong M, et al. (2008) AFM-based force spectroscopy measurements of mature amy-loid fibrils of the peptide glucagon. Nanotechnology 19:384013.

18. Kienle S, et al. (2014) Effect of molecular architecture on single polymer adhesion.Langmuir 30:4351–4357.

19. Krysiak S, Liese S, Netz RR, Hugel T (2014) Peptide desorption kinetics from singlemolecule force spectroscopy studies. J Am Chem Soc 136:688–697.

20. Adamcik J, et al. (2010) Understanding amyloid aggregation by statistical analysis ofatomic force microscopy images. Nat Nanotechnol 5:423–428.

21. Kad NM, et al. (2003) Hierarchical assembly of β2-microglobulin amyloid in vitro re-vealed by atomic force microscopy. J Mol Biol 330:785–797.

22. Sunde M, et al. (1997) Common core structure of amyloid fibrils by synchrotron X-raydiffraction. J Mol Biol 273:729–739.

23. Fändrich M, Meinhardt J, Grigorieff N (2009) Structural polymorphism of AlzheimerAbeta and other amyloid fibrils. Prion 3:89–93.

24. Sweers KK, Segers-Nolten IM, Bennink ML, Subramaniam V (2012) Structural modelfor α-synuclein fibrils derived from high resolution imaging and nanomechanicalstudies using atomic force microscopy. Soft Matter 8:7215.

25. Bauer HH, et al. (1995) Architecture and polymorphism of fibrillar supramolecularassemblies produced by in vitro aggregation of human calcitonin. J Struct Biol 115:1–15.

26. Kamgar-Parsi K, Hong L, Naito A, Brooks CL, 3rd, Ramamoorthy A (2017) Growth-incompetent monomers of human calcitonin lead to a noncanonical direct relationshipbetween peptide concentration and aggregation lag time. J Biol Chem 292:14963–14976.

27. Khalaf O, et al. (2014) The H50Q mutation enhances α-synuclein aggregation, se-cretion, and toxicity. J Biol Chem 289:21856–21876.

28. Ruggeri FS, et al. (2015) Influence of the β-sheet content on the mechanical propertiesof aggregates during amyloid fibrillization. Angew Chem Int Ed Engl 54:2462–2466.

29. Deng Z, et al. (2012) A close look at proteins: Submolecular resolution of two- andthree-dimensionally folded cytochrome c at surfaces. Nano Lett 12:2452–2458.

30. Dedmon MM, Lindorff-Larsen K, Christodoulou J, Vendruscolo M, Dobson CM (2005)Mapping long-range interactions in α-synuclein using spin-label NMR and ensemblemolecular dynamics simulations. J Am Chem Soc 127:476–477.

31. Garcia R, Perez R (2002) Dynamic atomic force microscopy methods. Surf Sci Rep 47:197–301.

32. Cerreta A, et al. (2012) FM-AFM constant height imaging and force curves: Highresolution study of DNA-tip interactions. J Mol Recognit 25:486–493.

33. Watson JD, Crick FHC (1953) Molecular structure of nucleic acids; a structure for de-oxyribose nucleic acid. Nature 171:737–738.

34. Ruggeri FS, et al. (2016) Nanoscale studies link amyloid maturity with polyglutaminediseases onset. Sci Rep 6:31155.

35. Ruggeri FS, et al. (2015) Infrared nanospectroscopy characterization of oligomeric andfibrillar aggregates during amyloid formation. Nat Commun 6:7831.

36. Smith JF, Knowles TP, Dobson CM, Macphee CE, Welland ME (2006) Characterizationof the nanoscale properties of individual amyloid fibrils. Proc Natl Acad Sci USA 103:15806–15811.

37. Bustamante C, Smith SB, Liphardt J, Smith D (2000) Single-molecule studies of DNAmechanics. Curr Opin Struct Biol 10:279–285.

38. Rief M, Gautel M, Oesterhelt F, Fernandez JM, Gaub HE (1997) Reversible unfolding ofindividual titin immunoglobulin domains by AFM. Science 276:1109–1112.

39. Bustamante C, Marko JF, Siggia ED, Smith S (1994) Entropic elasticity of lambda-phage DNA. Science 265:1599–1600.

40. Levine ZA, Larini L, LaPointe NE, Feinstein SC, Shea J-E (2015) Regulation and ag-gregation of intrinsically disordered peptides. Proc Natl Acad Sci USA 112:2758–2763.

41. Möglich A, Joder K, Kiefhaber T (2006) End-to-end distance distributions and intra-chain diffusion constants in unfolded polypeptide chains indicate intramolecularhydrogen bond formation. Proc Natl Acad Sci USA 103:12394–12399.

42. Rawat N, Biswas P (2014) Hydrogen bond dynamics in intrinsically disordered proteins.J Phys Chem B 118:3018–3025.

43. Fleming PJ, Rose GD (2005) Do all backbone polar groups in proteins form hydrogenbonds? Protein Sc 14:1911–1917.

44. Sheu S-Y, Yang D-Y, Selzle HL, Schlag EW (2003) Energetics of hydrogen bonds inpeptides. Proc Natl Acad Sci USA 100:12683–12687.

45. Fersht AR, et al. (1985) Hydrogen bonding and biological specificity analysed byprotein engineering. Nature 314:235–238.

46. Ragone R (2001) Hydrogen-bonding classes in proteins and their contribution to theunfolding reaction. Protein Sci 10:2075–2082.

47. Bolen DW, Rose GD (2008) Structure and energetics of the hydrogen-bonded back-bone in protein folding. Annu Rev Biochem 77:339–362.

48. Knowles TPJ, et al. (2007) Kinetics and thermodynamics of amyloid formation fromdirect measurements of fluctuations in fibril mass. Proc Natl Acad Sci USA 104:10016–10021.

49. Tsemekhman K, Goldschmidt L, Eisenberg D, Baker D (2007) Cooperative hydrogenbonding in amyloid formation. Protein Sci 16:761–764.

50. Knowles TP, et al. (2007) Role of intermolecular forces in defining material propertiesof protein nanofibrils. Science 318:1900–1903.

51. Zaccone A, et al. (2016) Kinetics of fragmentation and dissociation of two-strandprotein filaments: Coarse-grained simulations and experiments. J Chem Phys 145:105101.

52. Mannini B, et al. (2014) Toxicity of protein oligomers is rationalized by a functioncombining size and surface hydrophobicity. ACS Chem Biol 9:2309–2317.

53. Campioni S, et al. (2010) A causative link between the structure of aberrant proteinoligomers and their toxicity. Nat Chem Biol 6:140–147.

54. Fauvet B, et al. (2012) a-Synuclein in central nervous system and from erythrocytes,mammalian cells, and Escherichia coli exists predominantly as disordered monomer. JBiol Chem 287:15345–15364.

Ruggeri et al. PNAS | July 10, 2018 | vol. 115 | no. 28 | 7235

BIOPH

YSICSAND

COMPU

TATIONALBIOLO

GY

BIOCH

EMISTR

Y

Dow

nloa

ded

by g

uest

on

Janu

ary

24, 2

021